Messages in this thread | | | Subject | Re: [patch] real-time enhanced page allocator and throttling | From | Robert Love <> | Date | 06 Aug 2003 10:01:14 -0700 |
| |
On Wed, 2003-08-06 at 01:41, Andrew Morton wrote:
> It's pretty easy to demonstrate the benefit of the balance_dirty_pages() > change. Just do: > > while true > do > dd if=/dev/zero of=foo bs=1M count=512 conv=notrunc > done > > and also: > > rm 1 ; sleep 3; time dd if=/dev/zero of=1 bs=16M count=1 > > The 16M dd normally takes 1.5 seconds (I'm pretty please with that btw. > Very repeatable and fair). If you run the 16M dd with SCHED_FIFO it takes > a repeatable 0.12 seconds.
This is what I did. Same results, basically.
What I did not do was prove that the xmms stalls went away for those who were seeing that.
> So running a program off disk isn't a very good test.
No, its not. And in general, real-time tasks should not do disk I/O (at least not via their core RT thread). And they should mlock() their memory.
But circumstances do differ, and these changes are in the right direction, I think. It also means e.g. someone can make xmms or whatever real-time, and hopefully avoid the memory-related stalls that spawned the discussion and this patch.
Robert Love
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |